from research to internet standards · internet standards = ietf standards • the ietf is an open,...

Nokia Research Center

From Research to Internet Standards Lars Eggert Standardization Workshop Future Internet Conference Week Ghent, Belgium, 2010-12-15

Nokia Research Center 2

Who cares?


Researchers, why should you care about standards?

• If you’re researching Internet-related topics, where do you learn what the real current issues are?

• Hint: wireless ATM is not one of them

• You need to talk to operators, vendors, registrars, policy makers, regulators, etc.

• (Assuming you are interested in research that could have an actual impact)

• Where is it easy to meet these folks?

• Standards bodies + operator fora

3


But don’t forget to think for yourself

• You will talk to many folks who aren’t researchers • Their motivations are different than yours

• Often very short-term agendas • Few can abstract out to principles • Worried about the symptoms, not the causes • If all you have is a hammer, everything starts to look like a nail • Many are there to make money

(or keep others from taking theirs) • Think hard if the “problems” you learn about pass muster

• c.f. software engineering requirement documents

4


Still… go!

• If you’re interested in what the real problems are, you’ll get a good understanding by attending these fora

• If you’re interested in fixing some of them, you’ll need to participate more regularly

• Papers don’t get deployed

• For Internet and “Future” Internet stuff – due to SDO change control agreements – that means: participate in the IETF

• (3GPP, ITU-T, ETSI for special topics)

5


IETF participation takes time

• Standardization is very different from “fire & forget” academic publication/presentation venues

• The time commitment is substantial, both in terms of email discussion and meeting travel

• You will need to convince a diverse set of stakeholders of the value of your proposal

• Theoretically optimal ≠ practically optimal

• Business aspects and deployment incentives are critical (papers don’t get deployed)

• Don’t forget about the research arm – the IRTF

6


Need additional motivation?

• If you’re on an academic career path, standardization is unlikely to get you tenure

• But it doesn’t often hurt you either

• You will meet likeminded people to collaborate with

• And some of them have substantial budgets

• If you’re a junior researcher not on the academic career path, getting positively noticed in these fora can lead to an industry career…

7


IETF in a nutshell


Internet standards = IETF standards

• The IETF is an open, diverse and international community

• Network designers, operators, vendors, researchers, etc.

• Common goal: evolution of the Internet architecture and protocols & smooth operation of the Internet

• Participatory culture; open to anyone: people, not companies

• Produces Internet Standards (and other documents)

• It has a research arm – the IRTF

9


IETF by numbers

• 1-2000 people at 3 meetings/year

• from ca. 40-50 different countries

• Many, many more on mailing lists

• ~120 working groups (WGs)

• 8 Areas with 15 area directors (ADs)

• More than 6000 RFCs published

• More than 50000 Internet-Draft revisions submitted

• IRTF = ~12 research groups (RGs)

10

US

CN JP

SE

DE

FI

FR KR

Other

Participants at IETF-75 Stockholm, July 2009

1084 total, 50 countries


IETF standardization considerations

• Open process to produce open Internet standards

• Global standards for a global Internet

• Alignment with Internet architectural principles

• Maximum interoperability

• Maximum scalability

• Improved Internet security and privacy

11


IETF organizational structure


Top-level organizational view

RFC Editor

IAOC

IASA

IAD

IESG

Areas

13

“The IETF”


Top-level IETF & area structure

• IETF is structured into 8 areas • Each with area directors (ADs)

• Areas are structured into working groups (WGs)

• Each with WG chairs • Internet Engineering Steering

Group (IESG) = all ADs • Approves all Internet Standards • Manages technical work • Starts/ends WGs • Assigns WG Chairs

14

General Area

R. Housley

O&M Area

R. Bonica D. Romascanu

Security Area

T. Polk S. Turner

Internet Area

J. Arkko R. Droms

Routing Area

S. Bryant A. Farrell

RAI Area

G. Camarillo R. Sparks

Applications Area

A. Melnikov P. St.-Andre

Transport Area

L. Eggert D. Harrington

Internet Engineering Steering Group (IESG)

15 Area Directors

WG

WG

WG

WG

WG

WG

WG

WG

WG

WG

WG

WG

WG

WG

WG

WG

WG

WG

WG

WG

WG


Most active IETF organizations

15


IETF standards & documents


Internet-Draft (I-D) • Active working documents • Not finalized! Not stable! • Anyone can submit

• draft-yourname-... • Only some IDs are WG

documents! • draft-ietf-wgname-...

Request For Comment (RFC) • Archival publications

• Never change once published • Not all RFCs are Internet

standards! • Standards track =

Proposed/Draft/Full Standard • Other types =

Informational, Experimental, Best-Current-Practice (BCP)

17

IETF documents – two types


Origins of authors of recent RFCs

18


Origins of authors of recent Internet-Drafts

19


IETF document format

• English if the official language • ASCII is the mailing list and

document format • Frequent discussion of

alternate formats • IETF seen as “behind the

times” • (Almost) no drawings • But no consensus on

alternative • The current format is still

readable after 40+ years…

20


Bringing new work to the IETF


The IETF takes on work, when…

• There is a problem that needs solving • The problem fits one of the IETF areas

• Aligned with Internet architectural principles

• Scope is well defined and understood

• Research is complete, and engineering work is needed

• Agreement on specific deliverables

• Probability of timely completion

• People willing to do the work RFC

RFC editing process

IETF & IESG reviews

Working group process

Bring idea to the IETF

Idea & community building

22


Initiating New IETF Work – Existing WG

• Check WG charters & approach chairs to ask their opinion • Submit an I-D to the WG

• draft-yourname-wgname-topic-00 • Ask for feedback on I-D on WG mail list • Ask for presentation time during an IETF meeting • Constructively incorporate feedback

(“revise quickly, revise often”) • Eventually, ask to adopt as WG item • Continue work in WG (you now become editor)

23


WG, IETF and IESG Process

• Chair establishes WG consensus • Then requests publication of

I-D as RFC • I-D AD review by responsible AD • IETF-wide “Last Call” • IESG review

• Last Call comments & own technical review

• IESG approval • RFC editor process & publication

24

publication requested

Working Group

IESG

IETF Community

RFC Editor

IETF Last Call

comments, suggestions

significant issues

approval

RFC

I-D


Example: Better tools for IPv6 & IPv4 co-existence

• In 2008, service providers worried about the ability to deploy IPv6 fast enough (before IPv4 depletion)

• A series of bar, hallway and interim meetings led to a decision to develop some new technology for better co-existence in two WGs

• Results now complete; process took about 2 years

25

Individual I-D 248 days (32%)

WG I-D 263 days (33%)

IESG 205 days (26%)

RFC Ed. 59 days


Average time from Internet-Draft to RFC

26


Initiating New IETF Work – New WG

• Make sure no existing WG fits! • If “small”, can ask AD for I-D sponsorship • Else, likely need to organize a BOF (“Birds

of a Feather”) session at IETF meeting • Must form a community of interested

people around your proposal (!) • Read RFC5434 & prepare BOF proposal

• Problem statement I-D, open mailing list, draft BOF agenda, etc.

• Ask an AD for BOF sponsorship • BOF determines if a WG may form

27

may have BOF

chair, description, goals and milestones

IESG

Area Director

Working Group created

IAB

community


Example: PCN (Pre-Congestion Notification)

• Idea presented in TSVWG ca. 2005 • Bar meeting at IETF-66 in Dallas, TX Mar 2006 • PCN mailing list created Aug 2006 • draft-chan-pcn-problem-statement-00 posted Sep 2006 • First draft charter posted Sep 2006 • BOF requested Sep 2006 • BOF held at IETF-67 in San Diego, CA, USA Nov 2006 • Charter went for External Review Feb 2007 • WG chartered Mar 2007 • WG is ca. 50% done Dec 2010

28


Conclusion


Researcher • Hear about what the real

problems are • Work on meaningful open

issues – help build the Internet • Understand what promotes

and hinders deployment • Meet potential collaborators

and funding sources • Have a realistic understanding

of the time commitments

IETF • Gains highly skilled, unbiased

experts • Use academic results to create

better standards • Enable researchers to directly

improve the Internet • Insight into trends that will

impact standards down the road

• Accompany relevant topics in the IRTF research arm

30

Researcher participation in the IETF is important

from research to internet standards · internet standards = ietf standards • the ietf is an open,...

Documents